Hi,

I am in need for Regular Expression pattern that could help me parse what I would call tokens. Currently, I am parsing a string character by character and using a scanner approach (Sort of how compiler works). However, I would think a RegExp would perform better. 

Here is the scenario: Everything that is used within a “()” is called a token. A token could have nested tokens. An example would be:

Say, I have a token called (CurrentUser)and I have another token called (Date). Each token could support embedded attributes inside it. For example, the Date might have format=”Long|short|longtime|shortime”.

So if I have a string that is “(CurrentUser) logged on (Date: format=”short”). I should be able to replace that string by “John Smith logged on 03/30/3008” That was simple, however it could get more complex. For example say “(CurrentUser)  logged on (Date: format=”(defaultFormat)” ) . the way my algorithm works is first I find inner tokens

(CurrentUser) : John Smith

(defaultFormat) : Long

Step 1  “(CurrentUser)  logged on (Date: format=”(defaultFormat)” )

Step2  “(CurrentUser)  logged on (Date: format=”long” )

At this point, I start processing form start again and I will do string replacement on the tokens.

I am hoping to find a creative RegExp pattern that I could use to parse it in a better way.

Thanks,

This is not something a regex is well suited for. Regexes are for pattern matching, parsing is more complicated than that. Regexes are a poor substitute for a parser. You didn't mention which language you were doing this in but only a few regex engines support nested constructs. Even those won't return the nested value as a separate match  which is what I'm guessing you want.

---

Michael

"In theory, theory and practice are the same. In practice, they are not."
Albert Einstein